'Medical Moneyball' pitches winning health-informatics formula

Health informatics has the same potential to find game-changing medical treatments by sifting through patient data as the sports-loving computer geek who helped the lowly Oakland A's build a winning baseball team in "Moneyball," the Boston Herald reports.

"Medical Moneyball" is in play at Boston Children's Hospital, the newspaper reports, where Jonathan Bickel. M.D., is the director of clinical research informatics. The physician, who has a master's degree in informatics, and a team of eight software engineers help researchers analyze patient information to develop predictive models of healthcare, the paper reports.

"My dream is to take our huge database of patient information, with strict privacy controls in place, and create a moving toolkit to do large-scale data mining that looks at emerging trends and how patients with certain diseases are similar and different from one another," Bickel told the Herald.

Daniel Nigrin, Children's chief information officer, says the hospital's informatics specialists already are contributing data being used by public-health researchers for large-scale population disease studies.

Researchers and informatics specialists continue looking for ways to improve data mining and protect patient privacy while searching electronic health records.

Researchers at University College London recently published a study in BMC Medical Informatics & Decision Making about an algorithm that extracts data from free text in EHRs, reducing the need for time-consuming manual text review.

In Australia, Deloitte launched an online service earlier this month giving users access to locational statistical data on more than 50 health conditions and diseases including asthma, diabetes, mental health, arthritis and musculo-skeletal conditions, according to an article in Pulse IT Magazine.  

And in Cleveland, researchers recently found that using natural-language processing programs to  scrub clinical narrative text of personal health information is just as effective as human record annotation.

To learn more:
-read the Herald article
-see the Pulse IT report